Leveraging Machine Learning in the Mineral Discovery ProcessBy Antonio Gozain | Thu, 07/07/2022 - 14:13
Mineral exploration is not a new business, so many of the world’s more obvious deposits have been discovered years ago. In the current context with fewer, smaller and more expensive discoveries, the mining industry has turned to new technologies like machine learning (ML) to unlock the optimal economic potential of projects.
In October, GoldSpot Discoveries, a company leveraging ML to transform the mineral discovery process, announced its engagement with Ranchero Gold, the owner of the 22,267-ha Santa Daniela concession in Eastern Sonora. “We are thrilled to be working with GoldSpot. The application of ML to diverse exploration data is a new tool we can use to best identify new drill targets, and one of the best ways to move the exploration of new areas at Santa Daniela forward,” said William Pincus, Director and CEO, Ranchero Gold.
Previous operators of the Santa Daniela concession had evaluated and initiated drilling at Maiz Azul, El Rincon and Mesas Coloradas. Maiz Azul’s drilling resulted in the first gold discovery at Santa Daniela, explained Pincus. Since taking over the project in 2020, Ranchero Gold has conducted extensive exploration at both Maiz Azul and other areas of the concession block. This led to the company’s first drill program at Maiz Azul, which was completed in January 2022.
The Maiz Azul area is the most advanced prospect within the concession area. Recent drilling completed by Ranchero encountered gold in all 16 drill holes, including intercepts of 4.0g/t gold over 31.5m, 1.2g/t gold over 21.2m, 1.2g/t gold over 15.6m and 1.1g/t gold over 21m. The area lies at the northern end of the Santa Daniela concession block, within 3.5 kilometers of Alamos Gold’s Mulatos mine. “The best place to find a mine is still next to another mine,” shared Pincus.
GoldSpot Discoveries’ story began with six students participating in a challenge in Canada, explained Britt Bluemel, Senior Geochemist, GoldSpot Discoveries. While the first place went to a group of over 40 professionals, the students won the second place and were subsequently approached by impressed investors.
Since that moment in 2015, GoldSpot has grown to “a world-class team of over 80 people,” said Bluemel. Its team includes 34 geoscientists who specialize in field mapping, drilling, structural geology, geochemistry, resource estimation, 3D modeling and geophysics. The workforce also includes 25 data scientists, specialized in R&D, Artificial Intelligence, remote sensors, cloud computing, data analytics, back-end processing and front-end UX. “Our mission is to discover ore bodies more quickly, efficiently and with lower financial and environmental costs using technology,” said Bluemel.
ML is simply mathematics that look to provide answers. The real difference is made by geologists deciding whether the information provided is valuable or not, explained Bluemel. It is crucial to have geologists involved in the ML algorithm’s creation, curating the data layers to be sure that the outputs are related to relevant geological features.
The final objective of this technology’s implementation is to shorten the timeline between discovery and development, said Bluemel: “We want to predict the footprint of mineralization, to provide a score or probability to know where there are better chances to find the mineral and assign probabilities to certain areas.”